Unified depth prediction and intrinsic image decomposition from a single image via joint convolutional neural fields

Seungryong Kim, Kihong Park, Kwanghoon Sohn, Stephen Lin

Research output: Chapter in Book/Report/Conference proceedingConference contribution

26 Citations (Scopus)

Abstract

We present a method for jointly predicting a depth map and intrinsic images from single-image input. The two tasks are formulated in a synergistic manner through a joint conditional random field (CRF) that is solved using a novel convolutional neural network (CNN) architecture, called the joint convolutional neural field (JCNF) model. Tailored to our joint estimation problem, JCNF differs from previous CNNs in its sharing of convolutional activations and layers between networks for each task, its inference in the gradient domain where there exists greater correlation between depth and intrinsic images, and the incorporation of a gradient scale network that learns the confidence of estimated gradients in order to effectively balance them in the solution. This approach is shown to surpass state-of-the-art methods both on single-image depth estimation and on intrinsic image decomposition.

Original languageEnglish
Title of host publicationComputer Vision - 14th European Conference, ECCV 2016, Proceedings
EditorsBastian Leibe, Jiri Matas, Max Welling, Nicu Sebe
PublisherSpringer Verlag
Pages143-159
Number of pages17
ISBN (Print)9783319464831
DOIs
Publication statusPublished - 2016 Jan 1

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume9912 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Fingerprint Dive into the research topics of 'Unified depth prediction and intrinsic image decomposition from a single image via joint convolutional neural fields'. Together they form a unique fingerprint.

  • Cite this

    Kim, S., Park, K., Sohn, K., & Lin, S. (2016). Unified depth prediction and intrinsic image decomposition from a single image via joint convolutional neural fields. In B. Leibe, J. Matas, M. Welling, & N. Sebe (Eds.), Computer Vision - 14th European Conference, ECCV 2016, Proceedings (pp. 143-159). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9912 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-46484-8_9